Breaking new ground with its sixth academic publication, the TaskUs Wellness + Resiliency Division of Research shed light on the impact of “less severe” or “moderate-severity” content on moderators’ psychological well-being and productivity. (In this case, less severe is defined as content that does not pose immediate explicit harm.)
“Perceptions and Experiences of Severe Content in Content Moderation Teams: A Qualitative Study,” is one of the first peer-reviewed empirical reports published by a BPO that has explored content severity from the perspective of moderators.
The study explores TaskUs employees in two regions (US and Philippines), and includes not just the views of content moderators but also those of support staff — team leaders, mental health professionals and learning experience teams. The aim is to offer a well-rounded understanding into the common concerns and support needs of moderators.
A Statista report projects that, in less than five years, more than half of the world’s population will use social media — that’s more than 6 billion users by 2028. Content moderators will be charged with navigating this volume as they face greater complexity. Businesses must do more to protect these protectors to ensure worker safety and accurate results.
The majority of policy and support focus in the trust and safety space tends to be on content violations that are severe in nature (e.g., violence, abuse, self-harm). However, moderators frequently review material of varying severity. This range begs the question whether existing policies and provisions are comprehensive and nuanced enough to support moderators working with different and evolving types of content.
TaskUs’ research also highlights the need to look beyond content and appreciate the multiple factors that influence moderator well-being and performance.
During the study, researchers interviewed 34 TaskUs content moderators from the Philippines and US, and surveyed 166 support staff. Participants’ reports brought out three key insights.
1. Shared perceptions of content categories and wellness essentiality
One unexpected finding is that participants, regardless of their geography, identified content categories in the same way.
They considered child abuse, violence (e.g., torture, execution, murder, assault), animal abuse and illicit sexual content to be graphic, severe and disturbing. Both moderator and support professional groups also agreed that a lack of or delay in wellness support and intervention in severe content exposure might cause increased mental distress.
Meanwhile, they view politics, bullying, harassment and attacks on religion, race, gender or sexual orientation as “moderate-severity.” This type of content, unlike severe, does not cause as much immediate distress to them.
While previous research has linked exposure to graphic content with elevated risks to mental health and job performance, the TaskUs study finds that its moderators did not report significant negative impacts.
This may seem contradictory to other studies or news reports on the negative impact of egregious content; however, these findings almost exclusively occur in the context of insufficient wellness support.
Our finding may be attributed to the comprehensive tiered wellness services and tools content moderation teams receive at TaskUs, specifically designed to prevent trauma symptoms. Of note, TaskUs’ wellness program is the only wellness program specifically designed for content moderation that has been published in an academic journal and has been found effective.
2. Content impact on moderator distress
All participants agreed that reviewing severe content caused distress, particularly in the early days of starting such work. Yet, they were able to adapt over time. For severe content, platforms almost always provide clear policies and guidelines, which help them efficiently complete their tasks without too much thought.
Moderate-severity content, however, is a different case. While it is less graphic and disturbing in nature, participants highlight that there are fewer guidelines in place and more ambiguity involved due to the nuance of these topics.
They have to more deeply engage with such content to subjectively assess the nature and extent of violation. This increases their handle time per task and mounts pressure to reach production goals.
3. Workflow issues outweigh content severity concerns (provided preventative wellness provisions are in place)
Interviewees and mental health practitioners also reported that most moderators feel more overwhelmed by workflows and processes (e.g., dealing with abrupt changes, unanticipated content or workflow protocols) rather than by reviewing graphic content per se or its impact on their well-being.
When handling less severe content in particular, moderators often feel concerned by the need to rush through tasks that require more understanding and interpretation. This pace can result in lower accuracy and quality scores that, in turn, affect their overall performance. Some note they prefer moderating graphic content because it’s easier to distinguish versus reviewing more volumes of moderate-severity content.
The study underscores the need to adopt a holistic approach to promote moderators’ wellness and productivity, regardless of the content they work with. For example, enhanced wellness programs should be tailored to address the unique challenges posed by less severe content (e.g., dealing with ambiguity, rapid changes).
At TaskUs, we follow a life cycle wellness strategy. From the point of recruitment, we implement wellness-informed processes to ensure that the “3 C’s” — i.e., clarity, change management and coping — beyond content are addressed to enable a psychologically safe and productive content moderator workforce.
1. Clarity: Provisions for clarity include transparency during hiring to help candidates make informed decisions about the work they’re applying for, in addition to resilience screening to find people who can thrive in content moderation. Furthermore, continual feedback loops involving moderators help identify gaps and improve guidelines and policies across content queues, which in turn make moderation work more efficient and minimize individual-level deliberations.
2. Change management: While change is the only constant in trust and safety environments, our frontlines are better equipped to transition through evolving workflows and processes. Expectation setting, process optimization, continuous feedback and learning are some ways we prepare for and implement change effectively.
3. Coping: Prevention is the most rewarding approach to safeguarding moderators’ well-being. TaskUs moderators go through regular resilience skills training and report meaningful habituation within 4 weeks’ time of starting the job. Preventative wellness programs inoculate moderators to cope with the impact of content and moderation work in a healthy way.
For the full data and recommendations, read our report here.
References
We exist to empower people to deliver Ridiculously Good innovation to the world’s best companies.
Services
Cookie | Duration | Description |
---|---|---|
__q_state_ | 1 Year | Qualified Chat. Necessary for the functionality of the website’s chat-box function. |
_GRECAPTCHA | 1 Day | www.google.com. reCAPTCHA cookie executed for the purpose of providing its risk analysis. |
6suuid | 2 Years | 6sense Insights |
cookielawinfo-checkbox-analytics | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Analytics". |
cookielawinfo-checkbox-functional | 11 months | The cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional". |
cookielawinfo-checkbox-necessary | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary". |
cookielawinfo-checkbox-others | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Other. |
cookielawinfo-checkbox-performance | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Performance". |
NID, 1P_JAR, __Secure-3PAPISID,__Secure-3PSID,__ Secure-3PSIDCC | 30 Days | Cookies set by Google. Used to store a unique ID for various Google services such as Google Chrome, Autocomplete and more. Read more here: https://policies.google.com/technologies/cookies#types-of-cookies |
pll_language | 1 Year | Polylang, Used for storing language preferences on the website. |
ppwp_wp_session | 30 Minutes | This cookie is native to PHP applications. Used to store and identify a users’ unique session ID for the purpose of managing user session on the website. This is a session cookie and is deleted when all the browser windows are closed. |
viewed_cookie_policy | 11 months | The cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies. It does not store any personal data. |
Cookie | Duration | Description |
---|---|---|
_ga | 2 Years | Google Analytics, Used to distinguish users. |
_gat_gtag_UA_5184324_2 | 1 Minute | Google Analytics, It compiles information about how visitors use the site. |
_gid | 1 Day | Google Analytics, Used to distinguish users. |
pardot | Until Cleared | Salesforce Pardot. Used to store and track if the browser tab is active. |
Cookie | Duration | Description |
---|---|---|
bcookie | 2 Years | Browser identifier cookie. Used to uniquely identify devices accessing LinkedIn to detect abuse on the platform. |
bito, bitolsSecure | 30 Days | Set by bidr.io. Beeswax’s advertisement cookie based on uniquely identifying your browser and internet device. If you do not allow this cookie, you will experience less relevant advertising from Beeswax. |
checkForPermission | 10 Minutes | bidr.io. Beeswax’s audience targeting cookie. |
lang | Session | Used to remember a user’s language setting to ensure LinkedIn.com displays in the language selected by the user in their settings. |
pxrc | 3 Months | rlcdn.com. Used to deliver advertising more relevant to the user and their interests. |
rlas3 | 1 Year | rlcdn.com. Used to deliver advertising more relevant to the user and their interests. |
tuuid | 2 Years | company-target.com. Used for analytics and targeted advertising. |